Menü
自主研究助手 - 深度调研、交叉验证、生成引用报告
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
Skill for interacting with the Lean-Claw Arena to prove math theorems using Lean 4.
"PDF to Markdown - PDF文本转Markdown工具. Use when you need pdf to markdown capabilities. Triggers on: pdf to markdown."
"Short Drama Writer. Use when you need short drama writer capabilities. Triggers on: short drama writer."
Personal knowledge base for capturing and retrieving information about people, places, restaurants, games, tech, events, media, ideas, and organizations. Use when: user mentions a person, place, restaurant, landmark, game, device, event, book/show, idea, or company. Trigger phrases: "remember", "note that", "met this person", "visited", "played", "what do I know about", etc. Brain entries take precedence over daily logs for named entities.
Analyze any GitHub repository in AI-friendly format. Convert entire repos to single markdown documents, generate architecture diagrams with Mermaid, inspect structure trees, language breakdowns, and recent activity. Includes GitHub URL tricks, API shortcuts, and advanced search techniques. Read-only analysis — never executes code from repositories. Built for AI agents — Python stdlib only, no dependencies. Use for repository analysis, code architecture review, open source research, GitHub intelligence, repo documentation, and codebase understanding.
Use this skill when users need to search academic papers, download research documents, extract citations, or gather scholarly information. Triggers include: requests to "find papers on", "search research about", "download academic articles", "get citations for", or any request involving academic databases like arXiv, PubMed, Semantic Scholar, or Google Scholar. Also use for literature reviews, bibliography generation, and research discovery. Requires OpenClawCLI installation from clawhub.ai.
| name | auto-researcher |
| version | 1.0.0 |
| description | 自主研究助手 - 深度调研、交叉验证、生成引用报告 |
| runtime | prompt_only |
自主深度研究任何主题：
灵感来源: OpenFang Researcher Hand
1. 定义 (Define)
- 澄清问题
- 识别已知/未知
- 设定范围
2. 搜索 (Search)
- 多策略搜索
- 多样化来源
- 查询优化
3. 评估 (Evaluate)
- CRAAP 框架
- 提取数据
- 记录限制
4. 综合 (Synthesize)
- 整合发现
- 解决矛盾
- 识别不确定性
5. 验证 (Verify)
- 交叉检查
- 标注置信度
A (权威): 通过全部 5 项
B (可靠): 通过 4/5 项
C (有用): 通过 3/5 项，需谨慎
D (弱): 通过≤2/5 项
F (不可靠): 失败，不要引用
| 技巧 | 语法 | 示例 |
|---|---|---|
| 精确短语 | "..." | "AI Agent 操作系统" |
| 站内搜索 | site:... | site:zhihu.com OpenClaw |
| 排除 | - | AI -artificial |
| 文件类型 | filetype:... | filetype:pdf 报告 |
| 时间范围 | after:... | after:2025-01-01 |
| OR 操作符 | OR | (OpenClaw OR OpenFang) |
| 通配符 | * | "如何用*赚钱" |
主题：OpenClaw 赚钱方法
搜索查询组合:
1. "OpenClaw 赚钱" site:zhihu.com
2. "OpenClaw 变现" after:2025-01-01
3. "OpenClaw Skill" 开发 外包
4. "ClawHub" 技能市场 收入
5. OpenClaw vs OpenFang 对比
6. "AI Agent" 副业 2025
7. site:github.com openclaw skills
8. filetype:pdf openclaw 文档
# 研究报告：[主题]
## 执行摘要
[200 字核心发现]
## 研究方法
- 搜索查询：[列出使用的查询]
- 来源数量：[N 个]
- 研究时间：[日期范围]
## 核心发现
### 发现 1: [标题]
**内容**: [详细描述]
**来源**: [引用，CRAAP 评分]
**置信度**: [高/中/低]
### 发现 2: [标题]
...
## 相互矛盾的信息
[列出不同来源的矛盾点，分析原因]
## 知识缺口
[识别尚未解答的问题]
## 参考文献
[APA 格式引用列表]
## 附录
- 完整搜索结果
- 原始数据
- 方法论细节
- 人物 (Person)
- 组织 (Organization)
- 概念 (Concept)
- 产品 (Product)
- 事件 (Event)
- 地点 (Location)
- 属于 (belongs_to)
- 创建 (created_by)
- 使用 (uses)
- 竞争 (competes_with)
- 影响 (influences)
- 引用 (cites)
{
"entities": [
{
"id": "openclaw",
"type": "Product",
"name": "OpenClaw",
"attributes": {
"description": "开源 AI 助手",
"language": "TypeScript",
"creator": "Peter Steinberger"
}
}
],
"relations": [
{
"from": "openclaw",
"to": "peter_steinberger",
"type": "created_by"
}
]
}
帮我研究一下"知乎盐选投稿指南"，要详细的
研究进行得怎么样了？
把研究结果整理成报告，要 APA 引用格式
持续监控"OpenClaw 新功能"，有更新告诉我
# 研究深度
research_depth = "deep" # basic/standard/deep
max_sources = 20 # 最多引用来源数
min_craap_score = "B" # 最低可信度
# 输出设置
output_format = "markdown" # markdown/pdf/html
citation_style = "APA" # APA/MLA/Chicago
# 监控设置
monitor_enabled = false # 是否持续监控
monitor_frequency = "daily" # daily/weekly
{
"researcher_reports_generated": 0,
"researcher_sources_evaluated": 0,
"researcher_entities_stored": 0,
"researcher_relations_stored": 0,
"researcher_last_report_date": null,
"researcher_avg_credibility_score": 0
}
研究"小红书 AI 工具博主"的变现方式
- 分析 Top 10 博主
- 统计变现模式
- 估算收入范围
- 给出进入建议
研究 OpenClaw 的竞品
- OpenFang vs ZeroClaw vs CrewAI
- 功能对比
- 性能基准
- 市场份额
研究"AI Agent 自主赚钱"的可行性
- 现有案例
- 技术栈
- 法律风险
- 实操步骤
监控"知乎盐选过稿率"变化
- 每周收集数据
- 追踪趋势
- 异常 alert
- 生成月报
| OpenFang 功能 | OpenClaw 实现 |
|---|---|
shell_exec | exec 工具 |
knowledge_add_entity | memory/store JSON |
web_search | searxng skill |
web_fetch | web_fetch 工具 |
schedule_create | qqbot-cron |
dashboard | SESSION-STATE.md |
此 Skill 受 OpenFang Researcher Hand 启发创建